Prescient predicts which stories will matter to you in 24–72 hours, scores your forecast calibration as you read, and turns the daily news cycle into a game you actually want to play. Every AI prediction is tracked against 12 Brier-scored surfaces. Calibration is shown in the open Cockpit, not behind a sales deck.
12 Brier-tracked forecast surfaces predict topic momentum, decay rate, catalyst type, and reader trajectory — 4h to 7-day horizons. Every prediction is a falsifiable claim with a stated probability. At horizon resolution, the error is scored and published.
The Cockpit shows — in one screen — your AI forecast horizon, source diversity floor, doomscroll cap, and running Brier calibration score. Every dial is yours to tune. No hidden parameters.
Daily Word, Cluster Catch, Mini Crossword, and Headline Roulette are calibrated to the day's actual news. They're not themed to news — they're built from it. Every session contributes to your News IQ score.
Every prediction Prescient makes maps to one of these surfaces. At horizon end, the prediction is scored. The running per-surface aggregate is shown in the Cockpit. This is not a marketing claim — it's a testable substrate.
Prescient's CHARTER.md is public and immutable. Weakening any clause requires 90 days of public notice. The charter exists because the most-used news platforms have converged on engagement metrics and surveillance advertising. Prescient is a bet on the opposite premise: readers will pay for a product that respects them.
3 AI takes per day. 24-hour prescience horizon. One game archive day. Basic Cockpit. Every tier reachable.
Cockpit advanced panels. Unlimited AI takes. 72-hour prescience. Custom topic combos. Audio briefings. Prediction history export.
API access. Bulk RSS export. Calibration analytics export. Embedded prescience widgets. White-label briefings. Priority support.
Prescient is in beta. The 12 Brier-tracked surfaces are live and recording. Climatology values are bootstrapped priors until N≥30 per surface, at which point they will be refitted from observed data. The Cockpit shows the current N for each surface. We do not claim calibration is proven before the data says so.